Method of analysing DNA sequences

Abstract

The present invention relates to a method of identifying nucleic acid regions within a nucleic acid sample which interact with one another. In particular, the method relates to a chromatin conformation capture (3C) method which may be used to analyse the interactions between enhancers, silencers, boundary elements and promoters at individual loci at high resolution.

Description

CROSS-REFERENCE[0001]This application is a section 371 U.S. National phase of PCT / GB2016 / 053314, filed Oct. 24, 2016 which claims priority from GB 1518843.6, filed Oct. 23, 2015, which is incorporated by reference in its entirety.FIELD OF THE INVENTION[0002]The present invention relates to a method of identifying nucleic acid regions within a nucleic acid sample which interact with one another. In particular, the method relates to a chromatin conformation capture (3C) method which may be used to analyse the interactions between enhancers, silencers, boundary elements and promoters at individual loci at high resolution.BACKGROUND OF THE INVENTION[0003]Progress in our ability to annotate regulatory elements in the genome and determine their potential function has been driven by technological advances, such as RNA-seq [1], ChIP-seq [2, 3], DNase-seq [4] and ATAC-seq [5]. However, an outstanding challenge is to understand the mechanisms by which regulatory elements control specific gene...

AI Technical Summary

Benefits of technology

[0010]A further adaptation of the original Capture-C method is the Capture-Hi-C method (as exemplified in WO2015 / 033134) which combines the Capture-C method with the Hi-C library production. The Capture-Hi-C method requires the superimposition of a biotin group during the ligation step of library production to capture (as illustrated in FIG. 1 of WO2015 / 033134). This step is inefficient and so can greatly decrease the complexity of the library; ultimately, it greatly limits the sensitivity of the approach. Loss of complexity in the library is directly related to the amount of information that can be extracted per cell of the original sample. Therefore methods which involve such losses are only readily usable in circumstances where very large numbers of cells are available.

[0012]Analysis of the original Capture-C data showed that the sonication step allowed the identification of PCR duplicates in this assay, an effect that is problematic in most existing 3C methods. This allowed for the direct measurement of the efficiency of enrichment for ligation junction in the assay and made it possible to determine when all information had been extracted from a given library and where further sequencing would not yield any further information. By performing an experiment targeting a single gene promoter, it was surprisingly discovered that the captured DNA from this single region of interest made up less than 1% of the sequenced reads. This showed that great sequencing depth would be required in order to extract all of the information from the library with standard single capture. This would be completely impractical for smaller designs or repeated experiments, and it showed that simplifying the design itself would not readily increase the signal to noise ratio. It suggested that the signal to noise ratio was in fact inherent to the then current capture protocol.

[0015]Due to the complete control over PCR duplicates in the Capture-C protocol, it was realised that the library could be substantially over-amplified in the initial step of the library preparation part of the protocol, so that each informative junction would be represented multiple times prior to capture. It was further realised that, as junctions were now robustly represented and that the background was not intrinsic to the capture probes themselves, then the remaining background could substantially further mechanically-depleted by a second round of capture without loss of complexity.

[0016]The inventors have now found that the use of two sequential oligonucleotide capture steps applied to a 3C library (e.g. a PCR duplicated 3C library) prior to sequencing results in up to 3,000,000 fold enrichment compared to an uncaptured 3C library so that captured material now makes up approximately 50% (rather than 1%) of the sequenced material. This second capture step increases the number of PCR cycles and the number of PCR duplicates sequenced because the library complexity (i.e. the number of interactions available to capture) limits the number of unique interactions that could be sequenced. The greatly improved enrichment means that the depth of sequencing is no longer limiting. Using this new method, any PCR duplicates can be easily and efficiently excluded bioinformatically.

[0017]Furthermore, due to the huge increase in signal, independent 3C libraries (e.g. from different cell types or different stages of development) can now be captured and processed in a single tube making separately-indexed samples directly comparable. This greatly increases throughput and allows meaningful subtractive analysis of chromosome conformation in different cell types.

[0020]The investigation of gene regulation is not only limited by the number of genes or elements that can be interrogated, but also by the number of replicates, conditions, cell types and genetic variants that can be easily analysed. The huge increase in signal of NG Capture-C allows for the simultaneous capture of multiple samples in a single reaction, greatly increasing the throughput and economy of the assay. In practice, this allows complete networks of important genes, such as those encoding the Yamanaka pluripotency factors [32] (Myc, Sox2, Oct4, Klf4) to be analysed simultaneously in multiple cell types. The data are compatible with standard analytical tools and their reproducibility and comparability between active and inactive states of NG Capture-C provides a complementary approach to the statistical identification of regulatory elements. This complementary approach is capable of identifying all known regulatory elements at well characterised test loci, at levels of resolution previously not possible. Importantly, mindful of the current challenges in the analysis of GWAS and regulatory variants, the NG Capture-C method can been optimized to be effective at smaller cell numbers (approximately 100,000 cells) and to generate SNP-specific interaction profiles.

Problems solved by technology

The sheer variety of techniques which are currently available creates a challenge when attempting to improve yet further on the sensitivity of the basic 3C technique.

All 3C techniques require an enrichment step, the basis of which varies across most of the methods and it often unclear which will prove to be the most efficient and flexible approach.

Furthermore, issues such as the choice of restriction endonucleases or fragmentation method, cross-linking stringency, primer design, library complexity and probe position can all have an effect on the efficacy of any one particular enrichment method.

The original Capture-C protocol [12] used oligos synthesized on a microarray (Agilent SureSelect) with a minimum design of 40,000 oligos, irrespective of the number of desired viewpoints; the cost per sample was therefore very high for small designs.

Furthermore, the sensitivity possible with this previous design did not readily allow for the analysis of very long-range cis-interactions or trans-interactions and did not provide quantitative estimates of megabase-scale chromosomal interactions.

Importantly, although the method was high throughput at the level of which regions of the genome it could analyse simultaneously, it was still limited to one sample per assay.

This step is inefficient and so can greatly decrease the complexity of the library; ultimately, it greatly limits the sensitivity of the approach.

Loss of complexity in the library is directly related to the amount of information that can be extracted per cell of the original sample.

Therefore methods which involve such losses are only readily usable in circumstances where very large numbers of cells are available.

Images